Read between the numbers. Canada didn’t lose 42,000 jobs in August

Reporting on Statistics Canada’s latest Labour Force Survey (LFS), media outlets across the country claimed Canada “lost 42,000 jobs in August.” Many readers likely took this as evidence of U.S.-Canada tariff war damage. For others, ominous warnings about the effects of artificial intelligence (AI) likely came to mind. But the more important story in August’s figures has nothing to do with either.

The latest jobs report tells us that Canada’s labour force shrank by 36,800 people between July and August. After adjusting for seasonality, Canada’s labour force has not grown since October 2025. This is unprecedented outside of recessions. What explains it?

First, Canada’s aging population means more Canadians are entering retirement. Second, and likely more important in recent months, is the effect of the federal government’s policy U-turn on immigration and the number of newcomers entering Canada’s labour market every month. The population data back this up — Statistics Canada estimates show total population has been declining since July 2025.

It is also worth being precise about what the headline number actually measures. Despite its name, the data underlying the LFS do not count jobs, and Statistics Canada makes no claim that they do.

The LFS is a monthly survey of roughly 60,000 Canadian households, which asks respondents aged 15 and older whether they worked at a job or business last week. Statistics Canada then scales up the responses using population estimates to estimate the total number of employed people in Canada.

Between July and August 2026, that number decreased from 21,214,800 to 21,173,100 — a decline of 41,700 individuals. Did all of these individuals lose their jobs? No.

With fewer people in the country, it is no surprise that the number of employed people is also down. Last month’s employment decline of 41,700 overwhelmingly reflects what’s happening on the supply, not demand, side of Canadian labour markets.

But a lower employment level doesn’t necessarily tell us much about economic well-being — a point a recent C.D. Howe Institute report makes more extensively. It argues that weaker headline numbers simply become the norm as immigration slows and are not a sign of economic distress.

A bigger country isn’t necessarily a richer country. India has 30 times more employed workers than Canada, while Canada has seven times more than Norway. We surely aren’t aiming for an economy that is more like India’s and less like Norway’s.

What matters for economic well-being is the percentage of the population engaged in productive employment and the percentage of people who want a job but do not have one. To find evidence of tariff or AI effects going forward, we should focus on changes in employment and unemployment rates, not levels. Between July and August, Canada’s unemployment rate was unchanged at 6.4 per cent, while the employment rate inched down from 60.9 to 60.8 per cent.

At the same time, when reading monthly jobs reports, we should always remember that the numbers are estimates based on a sample of little more than 100,000 individuals in a population of 41 million people. Therefore, the estimates are inevitably noisy and the month-over-month changes volatile.

The expected error around the estimated July-to-August employment decline of 41,700 was ±33,500. At the conventional 95 per cent confidence level, the August decline is not statistically distinguishable from zero.

As our economic future becomes increasingly uncertain in the face of a trade war and rapid technological change, it is more important than ever to prioritize facts over easy narratives.

Mikal Skuterud, a fellow-in-residence and Roger Phillips Scholar of Social Policy at the C.D. Howe Institute, is a professor in the Department of Economics at the University of Waterloo.